Cluster Network Name

So is that to say there is downtime when a cluster fails over? or is that
stuff handled and there is just a "delay"?

Justin

"Edwin vMierlo [MVP]" wrote in
message
> It does not for W2003, as it does not have to:
>
> the IP address (in the resource group) goes offline, and comes online on
> the
> second node... it still is the same IP address, problem solved
>
> (this works differently when you are talking about W2008, multi-site, and
> on
> different IP subnets)
>
>
>
> "Balaji" wrote in message
>
>> Dear All,
>>
>> On a failover cluster, how does does windows redirect the traffic to
>> correct physical machine when the clients connect to the virtual
>> network name.
>>
>> Regards
>> Balaji
>
>
 >> Stay informed about: Cluster Network Name 

A cluster failover appears exactly like a service stop/restart to any client
applications. The client is reponsible for reconnecting and retrying any
failed transactions.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP





"Justin Rich" wrote in message

> So is that to say there is downtime when a cluster fails over? or is that
> stuff handled and there is just a "delay"?
>
> Justin
>
> "Edwin vMierlo [MVP]" wrote in
> message
>> It does not for W2003, as it does not have to:
>>
>> the IP address (in the resource group) goes offline, and comes online on
>> the
>> second node... it still is the same IP address, problem solved
>>
>> (this works differently when you are talking about W2008, multi-site, and
>> on
>> different IP subnets)
>>
>>
>>
>> "Balaji" wrote in message
>>
>>> Dear All,
>>>
>>> On a failover cluster, how does does windows redirect the traffic to
>>> correct physical machine when the clients connect to the virtual
>>> network name.
>>>
>>> Regards
>>> Balaji
>>
>>
>
>
 >> Stay informed about: Cluster Network Name 

Dear Edwin,

Are you saying the traffic is re-directed to the physical server that
is owning the cluster.

Regards
Balaji
On Feb 22, 3:22 pm, "Edwin vMierlo [MVP]"
wrote:
> It does not for W2003, as it does not have to:
>
> the IP address (in the resource group) goes offline, and comes online on the
> second node... it still is the same IP address, problem solved
>
> (this works differently when you are talking about W2008, multi-site, and on
> different IP subnets)
>
> "Balaji" wrote in message
>
>
>
> > Dear All,
>
> > On a failover cluster, how does does windows redirect the traffic to
> > correct physical machine when the clients connect to the virtual
> > network name.
>
> > Regards
> > Balaji
 >> Stay informed about: Cluster Network Name 

On Feb 22, 6:10 pm, "Justin Rich" wrote:
> So is that to say there is downtime when a cluster fails over? or is that
> stuff handled and there is just a "delay"?
>
> Justin
>
> "Edwin vMierlo [MVP]" wrote in
> messagenews:OHAMEzTdIHA.1132@TK2MSFTNGP06.phx.gbl...
>
> > It does not for W2003, as it does not have to:
>
> > the IP address (in the resource group) goes offline, and comes online on
> > the
> > second node... it still is the same IP address, problem solved
>
> > (this works differently when you are talking about W2008, multi-site, and
> > on
> > different IP subnets)
>
> > "Balaji" wrote in message
> >
> >> Dear All,
>
> >> On a failover cluster, how does does windows redirect the traffic to
> >> correct physical machine when the clients connect to the virtual
> >> network name.
>
> >> Regards
> >> Balaji

Hi Justin,

There is a downtime when the services hosted on the cluster goes
offline and comes online.

Regards
Balaji
 >> Stay informed about: Cluster Network Name 

>
> Are you saying the traffic is re-directed to the physical server that
> is owning the cluster.
>

Not sure what you are asking, but let me explain

In Windows 2003:
- The IP resource is moving with the Group to the other node, comes online,
after that it is the (R)ARP mechanism which will ensure the traffic is
received by the cluster node... standard IP networking (no "redirection")

In Windows 2008, and nodes on different subnets:
- There are two IP resources in the group, with an OR logical to the Network
Name, both IP resources will move to the node, but only one (for the correct
subnet) will come online, then the network name will come online. As now the
IP address has changed the "re-direct" (not really redirecting) is done by
DNS registration, so it relies on DNS to function properly. Same name in
DNS, IP address changed in DNS -> Clients can connect to the same name, new
IP address. Note: this might take some time to replicate the registration
around all DNS servers, as well as you might experience delay because of
local DNS caching.


Does that answer your question ?

HTH,
Edwin

PS, none of the two methods above are actually using "redirecting", it is
using standard IP networking technologies such as (R)ARP and DNS
registrations. Windows Failover Clustering is not really using any
"redirecting" technologies.
 >> Stay informed about: Cluster Network Name